[LV] Tail-folding, runtime scev checks

[SjoerdMeijer]

Summary

Now that we allow tail-folding, not only when we optimise for size, make
sure we do not run in this assert.

[Ayal]

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

2698–2699

Error message should be updated as well.

llvm/test/Transforms/LoopVectorize/X86/tail-folding-scevcheck.ll

13

Better also check instructions, not only labels?

[SjoerdMeijer]

Thanks, comments addressed

[SjoerdMeijer]

friendly ping

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

SCEVExpander Exp(*PSE.getSE(), Bypass->getModule()->getDataLayout(),
                    "scev.check");
   Value *SCEVCheck =
       Exp.expandCodeForPredicate(&PSE.getUnionPredicate(), BB->getTerminator());
 
   if (auto *C = dyn_cast<ConstantInt>(SCEVCheck))
     if (C->isZero())
       return;
 
-  // FIXME: FoldTailByMasking no longer implies OptForSize, this should be
-  // checking CM_ScalarEpilogueNotAllowedOptSize
-  assert(!Cost->foldTailByMasking() &&
-         "Cannot SCEV check stride or overflow when folding tail");
+  assert(!BB->getParent()->hasOptSize() &&
+         "Cannot SCEV check stride or overflow when optimizing for size");
 
   // Create a new block containing the stride check.
   BB->setName("vector.scevcheck");
   auto *NewBB = BB->splitBasicBlock(BB->getTerminator(), "vector.ph");
   // Update dominator tree immediately if the generated block is a
   // LoopBypassBlock because SCEV expansions to generate loop bypass
   // checks may query it before the current function is finished.
   DT->addNewBlock(NewBB, BB);

llvm/test/Transforms/LoopVectorize/X86/tail-folding-scevcheck.ll

+; RUN: opt < %s  -loop-vectorize -mcpu=core-avx2 -S | FileCheck %s
+
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+@a = external dso_local local_unnamed_addr global [1024 x i32], align 16
+
+define dso_local void @_Z5doit4ic(i32 %n, i8 signext %cstep) local_unnamed_addr #0 {
+; CHECK-LABEL: @_Z5doit4ic
+;
+; CHECK:       vector.scevcheck:
+; CHECK:       br i1 {{.*}}, label %scalar.ph, label %vector.ph
+;
+; CHECK:       vector.ph:
+; CHECK:       br label %vector.body
+;
+; CHECK:       vector.body:
+; CHECK:       call void @llvm.masked.store.v8i32.p0v8i32
+; CHECK:       br i1 {{.*}}, label {{.*}}, label %vector.body
+entry:
+  %conv = sext i8 %cstep to i32
+  %cmp9 = icmp sgt i32 %n, 0
+  br i1 %cmp9, label %for.body.preheader, label %for.end
+
+for.body.preheader:
+  %wide.trip.count = zext i32 %n to i64
+  br label %for.body
+
+for.body:
+  %indvars.iv = phi i64 [ 0, %for.body.preheader ], [ %indvars.iv.next, %for.body ]
+  %p.010 = phi i32 [ 0, %for.body.preheader ], [ %add, %for.body ]
+  %sext = shl i32 %p.010, 24
+  %conv1 = ashr exact i32 %sext, 24
+  %arrayidx = getelementptr inbounds [1024 x i32], [1024 x i32]* @a, i64 0, i64 %indvars.iv
+  store i32 %conv1, i32* %arrayidx, align 4
+  %add = add nsw i32 %conv1, %conv
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %exitcond = icmp eq i64 %indvars.iv.next, %wide.trip.count
+  br i1 %exitcond, label %for.end.loopexit, label %for.body, !llvm.loop !6
+
+for.end.loopexit:
+  br label %for.end
+
+for.end:
+  ret void
+}
+
+attributes #0 = { nofree norecurse nounwind uwtable }
+
+!6 = distinct !{!6, !7, !8}
+!7 = !{!"llvm.loop.vectorize.predicate.enable", i1 true}
+!8 = !{!"llvm.loop.vectorize.enable", i1 true}
+